Executive Summary

The U.S. Census Bureau maintains an important portfolio of economic statistics programs, including quinquennial economic censuses, annual economic surveys, and quarterly and monthly indicator surveys. Government, corporate, and academic users rely on the data to understand the complexity and dynamism of the U.S. economy. Historically, the Bureau’s economic statistics programs developed sector by sector (e.g., separate surveys of manufacturing, retail trade, and wholesale trade), and they continue to operate largely independently. Consequently, inconsistencies in questionnaire content, sample and survey design, and survey operations make the data not only more difficult to use, but also more costly to collect and process and more burdensome to the business community than they could be.

As part of an effort to reengineer its economic statistics portfolio, the Census Bureau requested a study by a panel of the National Academies of Sciences, Engineering, and Medicine’s Committee on National Statistics (CNSTAT) to review the annual economic surveys. The Panel on Reengineering the Census Bureau’s Annual Economic Surveys was charged to examine the design, operations, and products of 11 surveys and make recommendations to enable them to better answer questions about the evolving economy. The charge excluded the government sector (which was studied by a previous CNSTAT panel) and the agriculture sector (for which data are collected by the National Agricultural Statistics Service of the U.S. Department of Agriculture).

Aspects of the annual surveys that the panel was asked to consider included content, definition of the universe and data collection unit (e.g.,

company, establishment), sampling frame, sample design, questionnaire design, data collection, data processing, and estimation. The panel also was asked to consider improvements in data products of value to users with respect to accuracy, relevance, frequency, timeliness, and geographic detail.

AN INTEGRATED ANNUAL BUSINESS SURVEY SYSTEM

The panel’s overarching recommendation is for the Census Bureau to design and implement an integrated Annual Business Survey System (ABSS) to replace the current suite of separate annual surveys. The Census Bureau originally put forward this idea to the panel and users who were invited to discuss their data needs at an early panel meeting. After careful consideration, the panel concluded that an ABSS would make it possible to achieve a more efficient sample design and set of data collection, editing, estimation, and dissemination procedures, while at the same time improving the relevance, quality, and timeliness of the data collected and reducing the reporting burden on businesses. A unified business survey also would allow the Census Bureau to better manage its relationships with large, complex businesses, which typically receive more than one survey.

As part of the recommended ABSS, the panel endorses the use of small-area estimation models, which relate survey and administrative data to produce subnational estimates, a clear priority for many data users who spoke with the panel. The panel also recommends expanded use of administrative records and other information, as appropriate, to improve quality and timeliness and reduce burden for an ABSS.

• use administrative records to the maximum extent possible for greater efficiency and accuracy of data and reduced respondent burden; and
• incorporate small-area models to produce subnational geographic estimates with sufficient accuracy for users’ needs.

From design through implementation and continued operation of an ABSS, the Census Bureau should have ongoing programs to seek data user input and to evaluate and improve quality. Quality evaluation and improvement should encompass current and potential new data sources, data collection and processing, and resulting estimates.

IMPROVING EACH STEP OF THE SURVEY PROCESS

The panel analyzed the current annual economic surveys and makes specific recommendations for their improvement for each step in the design and conduct of a survey-based statistics program—keeping foremost in mind the goal of full implementation of an ABSS. The panel’s recommendations cover users’ needs, the sample frame, data content and collection, sampling and estimation, editing, imputation, disclosure control and quality standards, and data products and data access tools.

Users told the panel that they view Census Bureau economic data as the “gold standard” but require improvements in breadth, depth, timeliness, and accessibility. To ensure that the annual economic surveys—and, ultimately, the recommended ABSS—remain responsive to user needs in a changing economy, the Census Bureau will need to consult with them on a regular basis. This consultation should include explicit consideration of ways to minimize the potential effects of an ABSS on breaks in time series.

For the various aspects of the survey process, the panel’s recommendations include

• enhancing the Business Register to support its use as the sampling frame for all of the Census Bureau’s economic surveys and censuses and keeping it up to date;
• establishing account managers as the principal point of contact for large enterprises;
• maximizing the use of electronic data collection modes at all stages in the survey process;
• harmonizing core content, standardizing questionnaire format, and standardizing terminology and methods for sampling, editing, and other processing functions; and
• developing user-friendly data access tools.

The panel recognizes that it will take time to move from the current survey suite to an ABSS and commends the Census Bureau for the steps it has already taken toward survey integration. To be successful, the Census Bureau should assemble a team with sufficient authority to lead an ABSS development and identify initial activities for implementation that can build support for the project as a whole.